"Improvement" in the placebo group could be due to regression to the mean as well as to sociobiologic factors.